In my previous blog in this short series on generative AI in capital markets, I gave my perspective on why gen AI is so game-changing for our industry. From the highest level, using gen AI to drive efficiencies could help firms to respond to some of the bigger global macroeconomic challenges currently confronting the industry—not least inflationary and economic headwinds as these shifts may also demand firms to undertake cost transformation programs as part of what Accenture calls Total Enterprise Reinvention.
Also in the above-mentioned blog, I highlighted three groups of use cases where I believe gen AI could add significant value in capital markets:
- revenue generation
- customer focus
- operational productivity
Of these three sets of use cases, I believe that the first wave of gen AI projects will center mainly on the third: operational productivity.
Why is that? In a nutshell, the goal of any gen AI project is to augment the productivity of the human workforce by automating routine tasks and/or support employees with insights for decision making.
By tackling largely internally focused use cases, capital markets firms could gain experience and insight into what gen AI can deliver, without having too much direct involvement or impact on actual customer data and other potentially sensitive areas. That might make these projects, to a certain extent, a less risky place to start the gen AI journey.
Some use cases to target
So, as firms start seeking out efficiency opportunities opened up by gen AI, what use cases could they be targeting? Here are just a few of the many applications of gen AI that could offer a potential for efficiency gains in capital markets in the operations space.
One of the most interesting use cases is intelligent email routing. Whether email messages are from clients, counterparties or someone else, a gen AI engine can read and interpret them, and then route them automatically and instantaneously to the right person or team to handle them. It could also interpret queries to decide on the best next steps, and even draft replies for human employees to check and send back thus enhancing efficiency and reducing cost throughout.
Another area of operations where gen AI could deliver major efficiency gains is know-your-customer (KYC). Today, KYC processes such as checking and verifying a prospective client’s sources of funding are still largely manual—making them labor-intensive, costly, slow and error-prone. Gen AI could handle the complex checks required across various third-party bodies and databases at speed and generate granular reports detailing e.g., where their money is coming from. The outcome? Faster and better decision-making, and a far superior client onboarding experience at lower cost.
A third area where gen AI could probably help boost operational efficiency is around tackling fraud. Take payments fraud: Gen AI can monitor large transactional flows and identify and flag patterns that may be indicative of fraudulent activity. It can also reach out in human-like ways to a customer and check with them if a suspicious transaction that they’re undertaking is actually something they wanted to do. If the other party insists on pressing ahead with the transaction, the gen AI engine might escalate the case to a human specialist for a more detailed conversation.
Extending RPA end-to-end…
These are just a few of the use cases where gen AI can contribute to operational efficiency in capital markets. If you are searching for entry points within your company, I would suggest looking at areas where you have been leveraging robotic process automation (RPA) in the past. Why? Gen AI could extend and expand the use of robotic process automation (RPA) from a band-aid or workaround for specific problems, to an integrated solution that transforms entire processes end-to-end.
Business processes of all types could advance from partial to full automation, supported by unprecedented machine intelligence—driving further gains in efficiency and human productivity, while bringing employees new capabilities.
We might even see the potential of this tandem lead to an accelerated shift in sourcing models—a move “from offshore to botshore”. Firms could bring optimized and gen AI enabled processes back onshore, trying to link things together for even more efficiency gains.
…while committing to responsible AI—with a human always in the loop
All of this underlines why gen AI could bring such profound implications for capital markets firms’ operations. But to realize its full potential, firms need to keep two principles at the heart of their gen AI investments and implementations from my perspective:
- One is ensuring they always keep a human in the loop, mainly to check and further improve the quality and accuracy of any outcomes. And, of course, to keep oversight of the gen AI solution and process itself.
- The other is to commit fully to the concept of “responsible AI”, by ensuring their AI engines are secure, transparent, bias-free and respectful of people’s data privacy.
Clearly, this means complying with all relevant regulations in areas like privacy and confidentiality and also cyber security—all areas where gen AI could also help to pinpoint and close gaps. But it also requires firms to go beyond compliance in many ways. Like doubling-down on making data quality and integrity as high as possible. Testing AI outputs constantly and rigorously for signs of bias to address. And running “red hat” exercises to stress-test systems toward breaking-points and verify their resilience.
The big takeaway? In today’s fast-moving and cost-pressured capital markets industry, operational efficiency is key to success. And the next wave of transformational change in this area will be powered by gen AI in many areas. Firms should start moving now to harness it.
Thanks to my colleague Kevin Yang for contributing to this blog.